Object Vehicle Tracking by Convex Interpolation with Radar Accuracy

Abstract

In this paper, we propose a method to reduce the object vehicle tracking error of the radar system. A radar sensor is suitable as a vehicle sensor, but it has measurement uncertainty due to the operating mechanism and reflection points. To cope with these problems, we divide the zones uniformly and use relative location information to identify surrounding vehicles and then analyze the error characteristics of each zone. As a result of analyzing, the accuracy of the radar was different depending on the object vehicle position. Then, we applied a convex interpolation to consider radar accuracy. We used the convex interpolation method to effectively track the object. The resulting estimate is more accurate than the radar measurement. We confirmed the utility of the proposed method through a scenario-based object vehicle tracking experiment.